Key Frame Extraction Based On The Video Content And VLSI Implementation

With the rapid development of network technology and digital video technology, more and more information are stored and transmitted in the video format, now traditional methods for video retrieval can not meet the needs of users. How to manage and retrieval so large number of video data effectively and progressively becomes a hot research area. Shot boundary detection and key frame extraction is one of the two key technologies.The emphasis of our work focuses on the shot detection and key frame extraction algorithm under the video of YUV format. Firstly, we give a brief introduction and summarization of the existing shot boundary detection algorithm, and analyzed their advantages and disadvantages, in order to resolve the existing problems, we think out an algorithm based on sub-block histogram to detect abrupt change of the shot, while considering how to improve the threshold adaptability and the impact of removal of flash to avoid false detections. As for the shot gradual change, combined with the previous work and the analysis of the process of shot editing, we establish a mathematical model to simulate gradual process, using the method of linear regression gradients were detected. The results of the experiment show that the algorithm achieve a better results for the abrupt change; but to the gradual change, because we can not establish a model that can simulate all the gradual types, there stiil have much room the improve the performence of the algorithm. To improve the algorithm's performance, before the shot detection, we join the median filter to reduce the noise interfere to the video feature (luminance) of extraction.Then, after the shot segmentation, we consider how to extract the key frame to represent the main content of shot. In this method, we use the cumulate of the frame difference to represent changes of the content, the focuses of out discussion is on how to change the number of the key frame that we extracted on the length of the shot and reduce the number of redundant key frames.Finally, we discuss the algorithm and the RTL hardware implementation, and some of the key sub-module design was described in detail.